You can see a slide show of the retreating waters at Lake Mead and Hoover Dam and listen to my radio feature from The California Report. Also, The American Experience will rerun its documentary on Hoover Dam, Monday night on most PBS stations.

The Las Vegas Sun has a digital clock on its website, counting down to a theoretical doomsday when the city’s principal source of water would go dry. Wagering on that question may not have found its way into the sports books on the Strip–but it did become a lively pastime among engineers and hydrologists, when a report emerged from San Diego’s Scripps Institution, with a dire forecast. The paper, by climate physicist Tim Barnett, put the odds at 50-50 that Lake Mead, the giant reservoir behind Hoover Dam, would reach “dead pool” by 2017. That’s the point at which the dam shuts down and neither hydroelectric power nor water emerges from it.

The Barnett study “definitely raised eyebrows throughout the basin,” admits Terry Fulp, deputy director of the U.S. Bureau of Reclamation’s Lower Colorado Region, which operates Hoover Dam and Lake Mead. As it turns out, Barnett was a bit pessimistic. Subsequent work by him and others revealed that he overestimated the evaporation rate at Lake Mead, and omitted inflows below a certain point on the river.

The bottom line, according to Balaji Rajagopalan at the University of Colorado: Doomsday is not quite that near at hand. But that doesn’t mean it’s not on the horizon. “After 2027, the demand increase outpaces the supply decrease,” Rajagopalan told me in a recent interview. “And that’s why much of the risk explodes from 2027 to 2057.”

All of these studies are couched in probabilities, much in the same way that the Corps of Engineers talks about a “100-year” flood. Rajagopalan says: “Even in our study, we have a 50% risk [of dead pool], but that occurs in 2057. And that makes a big difference in terms of water managers, what they can do.”

One of those managers is Pat Mulroy, who directs the Southern Nevada Water Authority. Her constituents rely on Lake Mead for 90% of their water, so she says she’s not inclined to wait around for a consensus. “I mean, during the entire period of the ‘90s when we were bickering with our friends in the lower basin over surpluses, there was zero probability that the drought that we’re currently in was going to happen,” Mulroy told me.  “I’ve lost confidence in probabilities.”

The Bureau’s Fulp says the Colorado system leans heavily on the huge water storage capacity of Lake Mead and its sister reservoir upstream, Lake Powell. “We’ve known for decades that this system is highly variable and that’s why so much storage was built.” When filled to capacity (which it was, more or less, 10 years ago), Lake Mead alone can hold enough to put an area the size of Pennsylvania under a foot of water. But a 10-year drought has left Mead at just over 40% of capacity (so think of flooding something more the size of Costa Rica). Just as current evidence and climate models both point toward lessening flows on the Colorado, many parts of the southwest still see relatively high population growth.

Scientists continue to run their statistical models aimed at handicapping the Colorado’s demise as a dependable bringer of water. But as Fulp sums it up, “It’s really a debate about when. It’s not really ‘if.”

I regret an error of my own that appeared in the radio feature. I misstated the number of people in southern Nevada who are dependent on water from the Colorado. The correct number is about two million.